fix: increase timeout for name.publish and fix setup code

[alanshaw]

name.publish takes forever! I've increased the timeout to 5 minutes to accommodate.

I also moved the test (which adds a file for future tests) into the before block, I've removed the assertions that this test was making (there should be adequate tests for adding files elsewhere) and instead of hard coding the CID I've extracted it from the return value of the files.add call.

[victorb]

Could we instead of raising the timeout, make sure the daemon is started in offline mode for this test? In offline mode, the publish operation takes about 0.01 seconds, and gives the same output as-if it was online.

[alanshaw]

I'm guessing the name.resolve test wouldn't work if we did this?

[victorb]

I'm guessing the name.resolve test wouldn't work if we did this?

It would, as long as we're using the same daemon.

[alanshaw]

hmm, interesting bug, ipfs.name.publish after ipfs.id in offline mode causes error:

$ ipfs daemon --offline &
[1] 47216
$ Initializing daemon...
Swarm not listening, running in offline mode.
API server listening on /ip4/127.0.0.1/tcp/5001
Gateway (readonly) server listening on /ip4/127.0.0.1/tcp/8090
Daemon is ready
$ ipfs id
{
	"ID": "QmahispzgXbQ5ewaxSjdrst8KNLzREhVHdhxZwAUHQWqLb",
	"PublicKey": "CAASpgIwggEiMA0GCSqGSIb3DQEBAQUAA4IBDwAwggEKAoIBAQDC4Q6B6burDPTlfIkjlF72Q/G72GT7CDBzlZsGI9XKf3XCC5PQiTBcpvKEnUCoBvquiClWE1yJiM4bu1dH/YR5NDBlx9EQzDDmiVvS909cqokRHxA0NY5t83hs8gthoMt6wn15euG2e6LVohqefq6YiiFWgakxDcRyMclLWKiF2t7U6OhMMSd2YdHp15wcHac13ZmdNKhZ/TBhjeKowl4DjAixxLE0kmcK5MsABeQEA+BHukz425K3929onEriSkG7xkSrd0YwKOx6J2Apub7Df8yhcG1PFoTyfTUD/gxkYNrB2jjqUJaNXrK4QDbaDinBNS3KZUn4SQtmxzj2gi4PAgMBAAE=",
	"Addresses": null,
	"AgentVersion": "go-ipfs/0.4.14/5db3846",
	"ProtocolVersion": "ipfs/0.1.0"
}
$ ipfs name publish zdpuArHMUAYi3VtD3f7iSkXxYK9xo687SoNf5stAQNCMzd77k
Error: private key already loaded

I've added an issue here: ipfs/kubo#4978

I'm happy to change to using offline mode but it's not possible right now...we need the node ID so the tests can assert on the publish result.

[daviddias]

Thank you @alanshaw.

[victorb]

I'm really not a fan of having us use live IPNS in tests when it's as slow as it is. Can we investigate and see if providing the --local flag can work instead of --offline for now?

[daviddias]

Note it is not "live" as in "live network", the tests spin a mini IPFS network of 3 nodes.